This repository contains simple code to extract the following features from audio: Spectrogram, (Wavelet Transform) Scalogram and (Q Transform) Spectrogram. The code to extract the Spectrogram and (Q Transform) Spectrogram is available in Spectrogram.ipynb and the code to generate the scalogram is present in Scalogram.m.
The input must be a wav file.
For a sample input (audios/bogo.wav), the extracted features are present as qtrans.png, scalogram.png and spectrogram.png.
The presentation Speech Features.ppt briefly explains speech signals, features and noise.